Esther Du Pon

With T2DM under blood glucose medication, recommended self-management requires more than 2 hours every day [21]. Older adults, patients who are newly diagnosed, and those who are physically limited need even more time. Adhering to exercise and diet [21] recommendations are the most time-consuming tasks. Moreover, the results of self-management are not always immediately clear [22]. The use of an online platform would be an extra self-management activity and, as a result, patients could become overwhelmed.

To help patients start their first session on e-Vita, a training about the functionalities of the platform was organized. This training seemed to help patients to start their first session on the platform. During this training, patients were, among other things, supported by creating an account to obtain access to the platform. This task was considered complex and posed a considerable barrier to visiting the platform. Therefore, the given support was essential, especially for those who had not mastered the required computer skills. In case of problems, patients could contact a help desk, which 35 patients utilized. This service was provided by “Care Within Reach” (Stichting Zorg Binnen Bereik). Yet, this initiative again asked for active patient behavior.

We found that high eHealth literate patients used the platform more often than low eHealth literate patients. eHealth literacy can be defined as the ability to seek, find, understand, and appraise health information from electronic sources and apply the knowledge gained to addressing or solving a health problem [23]. There is a correlation between eHealth literacy and health literacy [24], which is defined as “the cognitive and social skills which determine the motivation and ability of individuals to gain access to, understand and use information in ways which promote and maintain good health” [25]. Lower health literacy is associated with poorer self-management skills [26]. In addition, health literacy may affect behaviors necessary for the development of self-management skills. Given that self-management strategies are core components for effective treatment of T2DM, low health literacy poses a considerable health concern. Because platform use is also associated with health literacy [27], platforms may not be suitable for older adults with lower eHealth literacy. There is a need to invest in developing online skills, particularly for older adults with lower eHealth literacy. Otherwise, these patients probably need other encouragements to help them manage their condition by means of an online platform.

Health outcomes
Finally, Chapter 6 reported the effects of PRISMA on patient and health outcomes. Results show that PRISMA did not improve knowledge, skills, and confidence for self-management; diabetes self-care behavior; health-related quality of life; and emotional well-being. These findings may be explained by the already high scores at baseline. There was limited room for improvement (ceiling effect) because the patients included in our study had generally well-managed T2DM. Possibly, in our sample, improvement of health outcomes was not necessary, or otherwise there is a need to focus on a specific subgroup of patients who have problems self-managing their disease. This group may benefit more from the advantages of group education. Because a combination of group and individualized interventions appear to be most effective [28], patients probably need additional encouragement to change their behavior, such as follow-up education in either group or individual settings. In addition, it was not possible to make a statement about the clinical effects. Unfortunately, the effects of group education tend to slowly decline over time [29]. This is common for most behavioral interventions, and thus, patients need periodic refresher courses to sustain the effects over time [28,30]. Patient motivation to sustain their newly learned self-management behaviors could be kept up with booster sessions or additional exposure to the course [31].

Studying patient care interventions in clinical practice

Methodological reflection
Our study has several strengths. First, our study had a randomized controlled trial design. We randomized the patients to an intervention or a control group. Randomization is considered to be the most powerful experimental design in clinical trials: by keeping other variables equal between groups any differences in outcome can be attributed to the intervention [32]. The primary advantage of randomization is the reduction of allocation bias, which derives from baseline variables that may influence outcome(s). Randomization ensures that baseline characteristics not known to be related to the outcome of interest are equally distributed among the groups [33]. We used block-randomization and the participants within each block were randomly assigned to treatment conditions. The second strength of our study was that the participants were randomized over all general practices, which reduced the risk of influences by general practice on the results. Third, we investigated several self-management–related outcomes: usage of an online care platform, self-efficacy, participation during the consultation, medication adherence, and health outcomes. This led to a rich data collection of platform log files, video recordings of patient-PN visits, pharmacy refill data, and health outcomes. Fourth, our study reflected daily practice, and the results found are ecologically valid and applicable in usual care.

However, because this pragmatic study was carried out in daily practice, it is impossible to have full control. Therefore, our study also has some limitations: (1) selection bias, (2) ethics of randomization, (3) low response rate or high drop-out rate, and (4) accuracy of adherence measures. First, selection bias was likely because the actual population with T2DM may be less healthy than the patients included in our study. Patients who experience difficulties with self-management behavior probably do not voluntarily take part in studies or show up at interventions. This could have made the results less generalizable to all patients with T2DM in the Netherlands. On the other hand, the characteristics of the patient sample (sex, age, education) did reflect the general Dutch population with T2DM. Second, an obstacle was receiving agreement from health care organizations to expose certain patients (i.e., those allocated to the control group) to a potential intervention later in time. When the study was introduced, participants seemed to be motivated to receive the intervention. The health care organizations’ concern was that this motivation could vanish after a couple of months. Our first plan was to follow the participants in the intervention group for 12 months, but in the end, all involved parties agreed to only 6 months. Therefore, we designed a randomized, two-arm, parallel-group, open-label trial of a 6-month duration with a 6-month extension period in which both groups received the intervention (wait-list control). Third, only 12% of the approached patients participated, and the drop-out rate varied from 41.8% to 88.8% for the clinical outcomes. This could have affected the generalizability of the study. For the clinical outcomes (registered annually during the routine diabetes checkups), we were dependent on the data delivered as part of usual care. Due to a large number of missing values, we decided not to report the clinical outcomes, as it was not possible to make a statement about the clinical effects. Unfortunately, imputation could not solve this high drop-out rate because then more data would be imputed than measured, making the data and results too unreliable. Possibly, patients were not interested in an online intervention, and moreover, such an intervention needs to be tailored more to patient preferences and abilities specifically. Fourth, adherence could have been measured more accurately compared to the MPR using electronic medication packaging (EMP) devices. EMP devices provide precise data on the manner in which the patient uses the medication. A widely used EMP device in medication adherence studies is the medication events system (MEMS). The MEMS had advantages over other adherence measures in providing continuous, reliable data on medication use [34]. It determines the precision with which the patient adheres to a prescribed diabetes regimen. The MEMS helps identify whether the nonadherence is sporadic, consistent, or any other abnormal medication-taking pattern. Although the accuracy of MEMS is undeniable, due to the large costs of the equipment, it is not the preferred measure in larger clinical trials such as our study [35].

Complex intervention
Several aspects of the study were considered complex. First, the intervention we studied meets the definition for complex interventions according to the Medical Research Council guideline [36]. The following factors make an intervention complex: a number of interacting components within the experimental and control interventions, number and difficulty of behaviors required by those delivering or receiving the intervention, a number of groups or organizational levels targeted by the intervention, a number and variability of outcomes, and the degree of flexibility or tailoring of the intervention permitted. An intervention with different components could also be (too) complex for patients. They were asked to attend three meetings, one of which was a platform training with computers. Handling the computer turned out to be difficult for most of the older participants. Therefore, the complexity of the intervention itself probably caused drop-outs. Third, T2DM is a complex condition. As mentioned before, T2DM differs from other chronic conditions by the relative demanding self-management requirements and self-management activities to control the condition on a daily basis. Therefore, patients could have become overwhelmed by a platform as extra self-management activity.

Furthermore, we used a relatively simple methodology for a complex phenomenon, which could have led to other limitations. PRISMA was embedded in routine primary diabetes care, so it was tested in the full spectrum of everyday usual care to maximize applicability. We took several steps to minimize the risk of a type III error, which occurs when evaluating a program that has not been adequately implemented [37]. Such low intervention reliability could decrease the interpretability of the collected data. For example, we had to account for different implementers (trainers), differences in intensity of training [33], and patients who did not attend PRISMA meetings. First, PRISMA was guided by different PNs and dieticians, and their work experience in the diabetes field varied. To compensate, all trainers used the same protocol and were trained to report any deviations from this protocol after the training; however, no deviations from the study protocol were reported. Second, differences in training intensity should have been prevented by the study design. Every topic of PRISMA was accurately registered in the protocol, including time and depth of discussion. However, the trainers had to anticipate and respond to their patients’ needs, which could have differed between trainings. This may have led to other questions and shorter or longer discussions about specific topics. Trainers reported that generally all topics were discussed, but the extent may have varied. Third, although 72% of the patients from the intervention group who agreed to participate in the study attended at least one meeting of PRISMA, the number of patients who visited both PRISMA meetings is unknown. Several people were involved in the process of monitoring the presence of the patients. This turned out to be too complex. In an ideal situation, one of the researchers or an assistant would personally attend all training sessions.

The implementation phase of the study can also be considered complex. During the implementation phase, a high degree of organization was expected from all participating parties. For example, the Medrie care group planned the PRISMA trainings, the PNs were responsible for the video recordings of their consultations, Care Within Reach acted as a platform helpdesk, and the researcher was responsible for coordinating all processes. Especially during the start of the study, all parties were enthusiastic and involved in the project. However, there were also difficulties. In a partnership, every party participates from its own perspective, and what works in practice is not always useful for research purposes and vice versa. To make decisions about practical issues, detailed discussions usually were needed. Over the months, we noticed better mutual understanding of each other’s points of view, and we became a team with the same mission: helping to improve the self-management skills of patients with T2DM. Together, we found an appropriate balance between the rules of science and the applicability in practice. We aspired to establish ‘’blended care’’ by using a small-scaled implementation strategy: only eight general practices were included. This helped us to focus on changing the attitude of caregivers toward active patient participation. The researcher was available for the PNs and visited them regularly, which the PNs highly appreciated; personal attention seemed the best way to enthuse general practices. In conclusion, in pragmatic clinical studies, it is simply not possible to completely control all variables. This should not be the goal either.

Inclusion
Despite our efforts to enthuse patients about the PRISMA program, only 12% of the approached patients participated. We can conclude that many patients are not interested in our intervention as offered, and the patients who attended the platform training barely used the platform later on. Some patients were persuaded by their HCP to participate in this project initially but later dropped out. This could have been prevented by involving the patients in terms of personal attention at the beginning of the project. We sent a package of information material about the project, which could have been too noncommittal. While this was indeed our intention because attending our project was voluntary, explaining the information as received and answering questions by telephone could have helped patients to decide to continue on the study. However, this procedure may be too time-consuming and, therefore, a challenge in usual care and can also affect the external validity.

Regardless of the lack of effects found in the current study, the fact that the majority of patients did not seem very interested in an intervention is enough of a reason to consider it ineffective in the given setting. However, our intervention is not an exception in diabetes care concerning a lack of effects. The lack of interest in diabetes education was also found by Gorter et al. (2010). They investigated preferences and opinions of patients with T2DM on education and found that most patients preferred not to receive diabetes education by means of a special diabetes course. More than 80% of the patients preferred to receive diabetes education during their regular diabetes check-ups. Only 3% preferred a special diabetes course [38]. However, we had expected that patients would be open to receive the PRISMA program because it enabled patients to discuss all aspects of their T2DM with trainers and other patients.

The literature also shows that many diabetes self-management education programs are underused, with a significant proportion of patients choosing to not attend despite classes being readily available [39,40]. In a recent systematic review [41], two broad categories of non-attenders were identified: (1) those who could not attend for logistical, medical, or financial reasons (e.g. timing, costs, or existing comorbidities) and (2) those who did not attend because they perceived no benefit from doing so, felt they had sufficient knowledge already, or had emotional and cultural reasons (e.g. no perceived problem, denial or negative feelings toward education). Moreover, the attenders seemed to be often those who already used the methods taught in the course [42]. Similarly, our sample consisted of patients who controlled their T2DM relatively well. This could have resulted in missing the more complex patients who need such a program most. In addition, the general practices who participated in our study probably are pioneers in diabetes care.

In our study, we applied the traditional opt-in methodology in which patients were asked to actively signal their willingness to participate. Recruitment requiring patients to opt in is associated with a sample of healthier participants [43] and a lower participation rate [43,44] than the opt-out method. In an opt-out methodology, patients are contacted repeatedly unless they signaled unwillingness to participate. Opt-out is generally well accepted by HCPs and patients and results in high participation rates [45,46]. Another option is studying the platform as integral (joint) component of the consultations. In our study, some of the PNs showed the platform to their patients during their consultations. They answered their patients’ questions or explained functionalities. However, the use of the platform was still noncommittal. When there is a need to use the platform as part of usual care (e.g. to enter or to check patient data), patients and PNs may become accustomed to it. However, it is probably a matter of time, and in a few years, eHealth and platforms may be part of usual care.

Implications for daily practice and future research
According to our study results, the following implications for daily practice and future research are suggested.

First, our study can inform HCPs and policymakers about how patients with T2DM can benefit from consultations with a PN. PNs should take into account the attitude and the potential lack of skills of their patients because these factors could affect the process of shared decision-making. Trusting relationships with their PNs, which are usually developed over time, help them to participate more actively [8]. Therefore, patients have a need for a PN “of their own.” Encouraging patients to write down questions in advance of the consultation or visiting the PN together with their spouses could also be valuable to participate actively. Our findings may provide evidence for designing interventions aimed at improving patient participation. In addition, as a result of the PRISMA program, HCPs should be prepared for a more counseling role in consultations with patients with T2DM. In further specifying patient goals of behavior change, they can encourage patients who are already engaged in self-counseling to improve their diabetes management. Future research should investigate the effect of PRISMA on consultations with PNs after six months. Patients may need more time to manage their health goals and participate more actively.

Second, the PRISMA program was originally developed with the purpose of increasing self-management behavior in patients with T2DM. However, HCPs and policy makers could take into account that medication adherence, as part of diabetes self-management, might be influenced by PRISMA. Our study was focused on suboptimal implementation of the dosing regimen, which is one of the phases of nonadherence. Since in T2DM nonadherence of the medication therapy mostly occurs during the initiation phase [47], HCPs should provide group education at an early stage (e.g. directly after diagnosing T2DM). Moreover, future research regarding adherence in patients with T2DM should focus on all three adherence phases: the initiation, the implementation, and the persistence of the pharmacotherapy. In addition, patients with T2DM are usually older adults in whom polypharmacy is common, and polypharmacy is a risk factor for slower initiation of medical treatment [47]. Therefore, this patient group needs extra attention. By doing so, patients might fully benefit from the advantages of group education. To date, the most recent version of PRISMA is extended with education about adherence. For future research, it would be of interest to test whether this version has an effect on adherence to prescribed medications and lifestyle recommendations.

Finally, in the future, due to the burden on diabetes health care, the care provided in general practice will be less intensive. However, it is important to uphold or improve the current level of quality of diabetes care. Therefore, it will be of interest to compare current usual care to a less intensive version of usual care (e.g., less contact moments between patient and PN) plus an intervention (e.g., an online care platform). Future research should also explore other sources for patients to develop intentions to behavior change for facilitating the use of self-management support programs within a platform. In addition, online care platforms may not be suitable for the older adults with lower eHealth literacy. There is a need to invest in developing online skills, particularly for this patient group. Otherwise, these patients probably need other encouragements to help them manage T2DM using an online platform.

Conclusion
This research aimed to investigate how self-management skills in patients with T2DM can be improved. We did not find an added value of the PRISMA program in improving the use of an online care platform. Our target group may have been uninterested in this type of intervention, which is the reality of usual care. It appeared to be difficult to engage patients in the PRISMA program, which resulted in a low inclusion rate and a large number of patients who did not attend (both) meetings. The management of T2DM already requires major effort and responsibility from patients, and group education may be considered yet another obligation. Therefore, other sources need to be explored for patients to develop intentions to behavior change for facilitating the use of an online care platform.

To deal with the increasing number of patients with T2DM and the burden of diabetes on health care, increased patient participation is needed, including more self-management. An improved online platform could be part of a solution. We observed that technical issues on the platform decreased patients’ motivation to further explore it, so the platform must be ready for use. In addition, a training to help patients use the platform seems essential. Older adults with lower eHealth literacy need extra support in using a platform. Moreover, we noticed a lack of enthusiasm in patients using the platform; however, embedding the platform in usual care is expected to make implementation easier. Nevertheless, in a few years eHealth and platforms could be part of usual care.

To date, the quality of clinical care for patients with T2DM in the Netherlands is already good. In the future, it is important to uphold this current level of quality of diabetes care or improve it with less intensive care from general practice (e.g., less face-to-face contacts with HCPs). Nevertheless, if the quality of care remains the same, we can regard it as profit.